Veera M. Rajagopal on using AI to pick content; Jerry Kaplan vs. doomerism; John Cochrane on the AI letter; Timor Kosters and others on an experiment with agents and humans
The shape of the near-future is beginning to become apparent: we will experience an AI tech-finance bubble, which will have a similar effect to the dotcom bubble of 27 years ago. Like the fiber-optic cable that was laid in the late-90's, the models and some of the architecture of AI will remain, to be built upon during the NEXT wave of AI advance.
Unless, that is, the AI's can figure out some way to overcome the energy and data center and corporate debt bottlenecks. Perhaps they already have...
General intelligence is one of the most studied phenomena in psychology...100 years and counting. Those steeped in Rousseau have yet to come close to disproving the phenomenon. It is just another human attribute that varies likely everything else. We all seem to be pretty good at distinguishing the bright from the dim.
I believe even Gardner admitted he was trying his luck labeling talents as "intelligences." I second the research saying learning styles are bunk.
The good news is the general public likes to read popular press articles about the latest news in psychology. The bad news is no one publicizes the meta-analyses twenty years later.
Every textbook I read in my teaching graduate program that broached the topic of intelligence endorsed multiple intelligences as settled fact. They didn't even call it a theory. That's fine insofar as it encourages teachers to differentiate their instruction beyond lecturing. I feel like the main reason it is used is to completely deny that g factor is real in favor of an unfalsifiable Myers-Briggs-style framework.
Recent research has indicated that people with different "learning styles" don't actually learn better if taught in their "style". "Auditory learners" don't learn better if taught orally. Visual learners don't learn better is taught using things they can see. Etc. Some subject matter is taught better orally or visually or through some other modality. But what is being taught determines how it is best taught, not some supposed "learning style" of the student.
I think it makes more sense to say that we all have different strengths and weaknesses, and in so many different categories, that talking about intelligence is something of a lost cause. We all know people who are "straight A's, can't tie their shoe laces". Some of the best and most illuminating conversations I have had were with carpenters and taxi drivers on subjects I never would have thought we had in common.
I don't intend this to be combative but your post illustrates one of the major problems with attempts to discuss the existence of general intelligence because it immediately gets equated with economic success and especially social status, in both directions.
It's difficult to analyze anecdotally because some individuals seem to be better examples of general intelligence or of singular domains within multiple intelligences when you examine their strengths and weaknesses. I know people like those you describe, but I also knew people growing up (at a prep school) who clearly had very high IQs and seemed to excel at everything they did with ease. Although I am sure there are people with high general intelligence who fall through the cracks in the education system and become blue-collar workers.
I am out of my depth here a bit, but I think g factor might be similar to how compute scales in AI models, where the more you increase it, the more improvements you see across all or nearly all domains.
This isn't related to the post, but something I have experienced with In My Tribe the last two weeks or so. When I try to read from the Subscriptions Tab on my iphone, the Substack app locks up. No problems with other subscriptions, just In My Tribe for some reason.
Anyone else run into that problem, here or elsewhere?
Please post an update if you find out the problem.
I usually start substack sessions by just typing “substack.com” in my laptop browser, which goes to the Home tab. Every once in a while, clicking Subscriptions or Activity pauses for 10+ seconds. One hung for over a minute but had completed when I came back five minutes later. I have found no pattern to it. Happens once or twice a week. I haven’t had any other hangups, such as specific substacks.
And speak of the devil! Trying to respond to your comment from the Activity tab with everything above was rejected with a complaint about attachments. Copying that and replying here, from the article's comments, will work, because this has happened before.
I get the weird delays and lag on my pc too from time to time. Just Wednesday night it was taking forever to load anything, and eventually I just gave up. Sometimes it runs perfectly, others it is a laggy mess. I assume it is because they are doing updates in the background that don't work right, but I don't know.
The thing about subscriptions locking on the phone is that it only has happened for In My Tribe, and it happens every time.
The problem with Veera M. Rajagopal's AI usage is she's only training it on its false positives -- rejecting articles which matched its filters. She's not training it on false negatives -- the articles it rejected which she might have wanted to read.
She's creating an echo chamber to confirm her bias.
"If general intelligence were everything, I could draw more than stick figures."
Of course it isn't everything. But it's further from nothing than everything. Intelligences are highly correlated. Not always, but in most cases.
As for your drawing, just because it doesn't come easy to you doesn't mean you couldn't have learned to be far better. And I'd bet that it coming easy isn't a requirement for being a great artist either.
Charles Murray dedicated a chapter or two to this topic in Human Diversity in the section about intelligence and class sorting. I feel like this is one of those things that leftist academics object to so deeply on a moral level that applying those insights in a mainstream way is very unlikely. East Asian countries don't really have those scruples as much and their education systems are actually closer to what Murray advocates.
Of course, I think people deep down do acknowledge this on some level, and that is why we still have things like magnet and exam schools (for now).
For some reason I've never been that curious about Murray's work. Perhaps because within my own life there are so many variables that aren't fixed, especially desire, which changes with maturity, reflection, and happenstance.
For example, I've been learning butterfly for over 25 years. I can swim a 50 well enough to get compliments from lifeguards. But when I honestly assess my practice, it hasn't been that good. So what's holding me back? Why am I not better?
What I want to understand is how my athletic intelligence and my analytical intelligence relate — where they're similar, where they diverge, and how correlated they actually are.
I immediately thought of the left's aversion to this topic and to Murray but then realized I didn't know for sure if he covered this aspect of intelligence.
I was also thinking the topic might not merit a book, or it was already written. Either way, a few research papers might be as useful.
The problem is not with them being Nobel laureates; it's with them thinking their particular Nobel expertise makes them experts in AI. They are appealing to authority.
Now if the were AI experts of some sort, that would be different. But there aren't 101 Nobel laureates in AI.
Do you trust Hollywood celebrities when they opine on unrelated fields? They must have above average intelligence to be so successful. Fauci must have had some intelligence to get his MD or PhD and rise so high. Elon Musk, Bezos, Gates, lots of smart successful CEOs. Their opinions outside their narrow expertise are meaningless.
Oh. but scientists, their intelligence is widespread in ways no other intelligence can match. Is that what you’re saying?
They are using their status as Nobel laureates to opine on topics completely unrelated to what earned them their respective prizes. Had they just signed their names without the laureate part, then all would be gucci in my mind. However, the “Nobel Laureate" designation was explicitly attached to their names in the petition, which is pretty cringe.
"I’ve never wanted to experience altered consciousness."
"Altered consciousness" is way overrated, and I have to question the motives of people who promote the use of psychedelics at any dosage. Speaking from a broad experience with those drugs in the 1960s, I can tell you that there is no enlightenment there. They scramble your brain. They muss up your perceptions. They can pull the rug out from under your sense of self and allow you to experience a form of insanity that can be terrifying -- but it's the terror of a dog suspended over an a painted abyss. Not to imply that the effects of these things are trivial, but the state they put you into has no more "spiritual" or intellectual content than spinning around to make yourself dizzy, pressing your eyeballs to create phospenes, or otherwise screwing up the normal workings of your mind and body to experience a novel sensation. Leave 'em alone, is my advice.
I would say that your recreational use of psychedelics tells us about as much about the potential benefits of these drugs as a young boy's use of firecrackers and cherry bombs tells us about beneficial use of explosives.
Thanks for the ideas Arnold. Here’s a little sci-fi mini post inspired by your first link.
The Weakness Feed
I still choose the weaknesses I want to work on—the ones I can see, in my writing, my thinking, my body, my marriage, my game—though the AI now suggests others it notices, and I decide which make the list. Imagine the agents doing the rest from a richer feed: a small wearable camera and microphone that captures everything around me from the moment I wake—what I see, what I say, what I hear, how I move—through the morning pages, the pickup game where my crossover breaks down under pressure, every conversation, every argument, even the most private hours of a marriage. They ingest it all through sealed, encrypted processing—no video kept, no audio kept, no voices, no names, nothing anyone could ever retrieve—apply my weakness filter, and present a curated list of readings on a dashboard that I try to visit every day. I give my feedback on the curation, marking each item as one of the following: love, like, good to know, or skip. The system takes that feedback, weighs it against what the device witnessed and discarded, and gets a little better at finding my blind spots each cycle. Once a week it picks around eight readings, often clustered around a single weakness—a point guard's memoir, a book on marriage I'd never have pulled off the shelf—and drafts them into a newsletter. I do not write that prose by hand, though I keep nudging it towards more clarity and brevity, and towards the flaws I would never have named myself. I do all of this for a simple reason: everyone else already does, and I don't intend to fall behind.
The acceptance of one’s alleged weaknesses might be more powerful than the infinite treadmill. In other words, staying present to those that mean the most to you is probably more important than the never ending quest for self-improvement.
If you are able to provide for your family and provide them with a supportive and loving environment, then I’d suggest that you have accomplished everything that is required of yourself. Anything beyond this seems like a “first world problem,” which is more likely to entrap than to provide sustained happiness.
The shape of the near-future is beginning to become apparent: we will experience an AI tech-finance bubble, which will have a similar effect to the dotcom bubble of 27 years ago. Like the fiber-optic cable that was laid in the late-90's, the models and some of the architecture of AI will remain, to be built upon during the NEXT wave of AI advance.
Unless, that is, the AI's can figure out some way to overcome the energy and data center and corporate debt bottlenecks. Perhaps they already have...
https://jmpolemic.substack.com/p/the-first-bubble-of-the-age-of-hyper
General intelligence is one of the most studied phenomena in psychology...100 years and counting. Those steeped in Rousseau have yet to come close to disproving the phenomenon. It is just another human attribute that varies likely everything else. We all seem to be pretty good at distinguishing the bright from the dim.
I believe even Gardner admitted he was trying his luck labeling talents as "intelligences." I second the research saying learning styles are bunk.
The good news is the general public likes to read popular press articles about the latest news in psychology. The bad news is no one publicizes the meta-analyses twenty years later.
Every textbook I read in my teaching graduate program that broached the topic of intelligence endorsed multiple intelligences as settled fact. They didn't even call it a theory. That's fine insofar as it encourages teachers to differentiate their instruction beyond lecturing. I feel like the main reason it is used is to completely deny that g factor is real in favor of an unfalsifiable Myers-Briggs-style framework.
Recent research has indicated that people with different "learning styles" don't actually learn better if taught in their "style". "Auditory learners" don't learn better if taught orally. Visual learners don't learn better is taught using things they can see. Etc. Some subject matter is taught better orally or visually or through some other modality. But what is being taught determines how it is best taught, not some supposed "learning style" of the student.
I think it makes more sense to say that we all have different strengths and weaknesses, and in so many different categories, that talking about intelligence is something of a lost cause. We all know people who are "straight A's, can't tie their shoe laces". Some of the best and most illuminating conversations I have had were with carpenters and taxi drivers on subjects I never would have thought we had in common.
I don't intend this to be combative but your post illustrates one of the major problems with attempts to discuss the existence of general intelligence because it immediately gets equated with economic success and especially social status, in both directions.
You're going to have problems discussing general intelligence if you insist in ignoring economic success and social status.
It's difficult to analyze anecdotally because some individuals seem to be better examples of general intelligence or of singular domains within multiple intelligences when you examine their strengths and weaknesses. I know people like those you describe, but I also knew people growing up (at a prep school) who clearly had very high IQs and seemed to excel at everything they did with ease. Although I am sure there are people with high general intelligence who fall through the cracks in the education system and become blue-collar workers.
I am out of my depth here a bit, but I think g factor might be similar to how compute scales in AI models, where the more you increase it, the more improvements you see across all or nearly all domains.
This isn't related to the post, but something I have experienced with In My Tribe the last two weeks or so. When I try to read from the Subscriptions Tab on my iphone, the Substack app locks up. No problems with other subscriptions, just In My Tribe for some reason.
Anyone else run into that problem, here or elsewhere?
I tried just now; Android, Samsung S23 Ultra. No problems.
Thanks for the confirmation. Hopefully this is just my phone being stupid and not something with Arnold’s blog.
Please post an update if you find out the problem.
I usually start substack sessions by just typing “substack.com” in my laptop browser, which goes to the Home tab. Every once in a while, clicking Subscriptions or Activity pauses for 10+ seconds. One hung for over a minute but had completed when I came back five minutes later. I have found no pattern to it. Happens once or twice a week. I haven’t had any other hangups, such as specific substacks.
And speak of the devil! Trying to respond to your comment from the Activity tab with everything above was rejected with a complaint about attachments. Copying that and replying here, from the article's comments, will work, because this has happened before.
Will do man.
I get the weird delays and lag on my pc too from time to time. Just Wednesday night it was taking forever to load anything, and eventually I just gave up. Sometimes it runs perfectly, others it is a laggy mess. I assume it is because they are doing updates in the background that don't work right, but I don't know.
The thing about subscriptions locking on the phone is that it only has happened for In My Tribe, and it happens every time.
”I’ve never wanted to experience altered consciousness."
But have you?
Drunk? Buzzed? High fever? Loopy drugs for a colonoscopy? Other disorienting prescription drugs?
The problem with Veera M. Rajagopal's AI usage is she's only training it on its false positives -- rejecting articles which matched its filters. She's not training it on false negatives -- the articles it rejected which she might have wanted to read.
She's creating an echo chamber to confirm her bias.
"If general intelligence were everything, I could draw more than stick figures."
Of course it isn't everything. But it's further from nothing than everything. Intelligences are highly correlated. Not always, but in most cases.
As for your drawing, just because it doesn't come easy to you doesn't mean you couldn't have learned to be far better. And I'd bet that it coming easy isn't a requirement for being a great artist either.
Intelligences are highly correlated. Huge, important topic. There's a book opportunity here for someone.
Charles Murray dedicated a chapter or two to this topic in Human Diversity in the section about intelligence and class sorting. I feel like this is one of those things that leftist academics object to so deeply on a moral level that applying those insights in a mainstream way is very unlikely. East Asian countries don't really have those scruples as much and their education systems are actually closer to what Murray advocates.
Of course, I think people deep down do acknowledge this on some level, and that is why we still have things like magnet and exam schools (for now).
For some reason I've never been that curious about Murray's work. Perhaps because within my own life there are so many variables that aren't fixed, especially desire, which changes with maturity, reflection, and happenstance.
For example, I've been learning butterfly for over 25 years. I can swim a 50 well enough to get compliments from lifeguards. But when I honestly assess my practice, it hasn't been that good. So what's holding me back? Why am I not better?
What I want to understand is how my athletic intelligence and my analytical intelligence relate — where they're similar, where they diverge, and how correlated they actually are.
I immediately thought of the left's aversion to this topic and to Murray but then realized I didn't know for sure if he covered this aspect of intelligence.
I was also thinking the topic might not merit a book, or it was already written. Either way, a few research papers might be as useful.
The idea that if a 101 Nobel prizewinners sign something, it must necessarily be stupid, is a peculiar sort of categorical declaration.
The problem is not with them being Nobel laureates; it's with them thinking their particular Nobel expertise makes them experts in AI. They are appealing to authority.
Now if the were AI experts of some sort, that would be different. But there aren't 101 Nobel laureates in AI.
Ah, like non prizewinning economists expounding on biology or ecology. Got it.
So g is not a thing after all - or not among Nobel laureates? They are more like autistic people with one quirky expertise?
Do you trust Hollywood celebrities when they opine on unrelated fields? They must have above average intelligence to be so successful. Fauci must have had some intelligence to get his MD or PhD and rise so high. Elon Musk, Bezos, Gates, lots of smart successful CEOs. Their opinions outside their narrow expertise are meaningless.
Oh. but scientists, their intelligence is widespread in ways no other intelligence can match. Is that what you’re saying?
Back to multiple, discrete, walled off intelligences?
No, back to asking why scientists, and especially Nobel laureates, are the only ones credited with general expertise outside their field.
They are using their status as Nobel laureates to opine on topics completely unrelated to what earned them their respective prizes. Had they just signed their names without the laureate part, then all would be gucci in my mind. However, the “Nobel Laureate" designation was explicitly attached to their names in the petition, which is pretty cringe.
"I’ve never wanted to experience altered consciousness."
"Altered consciousness" is way overrated, and I have to question the motives of people who promote the use of psychedelics at any dosage. Speaking from a broad experience with those drugs in the 1960s, I can tell you that there is no enlightenment there. They scramble your brain. They muss up your perceptions. They can pull the rug out from under your sense of self and allow you to experience a form of insanity that can be terrifying -- but it's the terror of a dog suspended over an a painted abyss. Not to imply that the effects of these things are trivial, but the state they put you into has no more "spiritual" or intellectual content than spinning around to make yourself dizzy, pressing your eyeballs to create phospenes, or otherwise screwing up the normal workings of your mind and body to experience a novel sensation. Leave 'em alone, is my advice.
I would say that your recreational use of psychedelics tells us about as much about the potential benefits of these drugs as a young boy's use of firecrackers and cherry bombs tells us about beneficial use of explosives.
What's my drug of choice?
Well, what have you got?
I don't go broke
And I do it a lot
RIP Layne Staley
https://youtu.be/p9GAEFTeWko?si=fvTVQy0ZTuJQKz5q
Thanks for the ideas Arnold. Here’s a little sci-fi mini post inspired by your first link.
The Weakness Feed
I still choose the weaknesses I want to work on—the ones I can see, in my writing, my thinking, my body, my marriage, my game—though the AI now suggests others it notices, and I decide which make the list. Imagine the agents doing the rest from a richer feed: a small wearable camera and microphone that captures everything around me from the moment I wake—what I see, what I say, what I hear, how I move—through the morning pages, the pickup game where my crossover breaks down under pressure, every conversation, every argument, even the most private hours of a marriage. They ingest it all through sealed, encrypted processing—no video kept, no audio kept, no voices, no names, nothing anyone could ever retrieve—apply my weakness filter, and present a curated list of readings on a dashboard that I try to visit every day. I give my feedback on the curation, marking each item as one of the following: love, like, good to know, or skip. The system takes that feedback, weighs it against what the device witnessed and discarded, and gets a little better at finding my blind spots each cycle. Once a week it picks around eight readings, often clustered around a single weakness—a point guard's memoir, a book on marriage I'd never have pulled off the shelf—and drafts them into a newsletter. I do not write that prose by hand, though I keep nudging it towards more clarity and brevity, and towards the flaws I would never have named myself. I do all of this for a simple reason: everyone else already does, and I don't intend to fall behind.
The acceptance of one’s alleged weaknesses might be more powerful than the infinite treadmill. In other words, staying present to those that mean the most to you is probably more important than the never ending quest for self-improvement.
I readily admit my weaknesses, but those that know me best would prefer change. What's this about staying present?
If you are able to provide for your family and provide them with a supportive and loving environment, then I’d suggest that you have accomplished everything that is required of yourself. Anything beyond this seems like a “first world problem,” which is more likely to entrap than to provide sustained happiness.